Do-not-call controls for AI campaigns: a practical review

A contact list is not permission to call. Before an AI campaign starts, the business needs to know why each contact is eligible, which restrictions apply and how a new opt-out stops future attempts. Those decisions belong in a reviewed contact policy and durable records, not only in the assistant's prompt.

Establish the applicable rules

The FTC's Telemarketing Sales Rule compliance guide describes federal telemarketing requirements and explains that state law can also apply. Registry screening, company-specific requests, consent and calling practices are related but distinct questions.

The FCC's 2024 declaratory ruling confirms that AI-generated human voices fall within the TCPA's artificial or prerecorded voice restrictions. Do not assume a conversational model creates an exemption. Have the actual call purpose, recipients, jurisdictions and consent process reviewed before launch rather than relying on a general blog article as legal clearance.

Maintain one authoritative contact decision

Store the source and scope of the permission relied upon, relevant restrictions and subsequent changes. A CRM label such as “lead” or “customer” is not enough to explain eligibility. Identify who owns the decision when a source record is incomplete or contradictory.

Registry checks and internal suppression records have different roles. Record the evidence your policy requires for each. Protect these records from casual import overwrites and make the reason for a blocked attempt visible to the campaign operator.

Test the race conditions

A campaign may be scheduled hours before it dials. A contact can withdraw permission after import, after scheduling or while another attempt is queued. Evaluate these cases with controlled numbers:

1. Suppress a contact after scheduling and confirm it is not dialed.
2. Issue an opt-out during a test call and verify the durable record.
3. Re-import the same contact and confirm the suppression survives.
4. Retry a failed job and confirm it rechecks eligibility.
5. Run another campaign against the same contact and verify that the relevant restriction remains effective.

These are acceptance tests to perform, not claims that every campaign product already passes them.

Make opt-outs an operational path

Decide how spoken requests, messages and staff-entered requests reach the authoritative suppression record. Confirm failure handling when that write is unavailable. A polite response from the assistant is insufficient if the next job can still call the number.

Retain the required evidence according to the organization's reviewed policy. Avoid unnecessary sensitive information in logs; access and retention need deliberate choices too.

What to verify in Burki

Burki's optional campaign and contact features do not amount to a managed legal-compliance service. Do not assume automatic national-registry subscriptions, universal state-law checks or infallible opt-out recognition. Confirm the controls available for the specific carrier, runtime and campaign before using a real list.

Start with a bounded internal test, a reviewed caller script and an explicit spend limit. If the contact policy or suppression path is unresolved, resolve that first. A successful AI conversation does not establish that placing the call was permitted.